A central goal in neuroscience is to identify cellular ensembles supporting mental and behavioral states, but these ensembles cannot be defined a priori. The dentate gyrus (DG), for example, contains more than 1M granule cells, which are essentially indistinguishable from each other, but less than 5% of these seemingly identical neurons are active during any one behavioral event, suggesting that the associated mental states are each mediated by a small subset of neurons. We propose to develop a novel method for identifying and gaining genetic access to such transient, behaviorally-relevant assemblies of neurons in awake animals. The key unique features of our approach are (1) its temporal precision is unprecedented because it is the first neuronal tagging technique that matches the timescale of naturalistic behavior; and (2) its ability to label multiple cell populations in the sme animal enables the comparison of state-specific cell ensembles. Our novel molecular-genetic technique first identifies activated neurons on the basis of elevated intracellular calcium and then tags them using light. Light application is especially attractive because it is temporally precise: just as other optogenetic methods have aided neuronal circuit analysis by approximating the timescale of cell activity, so too will a light-dependent labeling technique illuminate functional cell assemblies. The technique will be entirely virus-based, so it is usable across species without relying on transgenic animals. Under this award we will establish the technique by developing and testing two critical innovations: (1) a synthetic bidirectional promoter system, and (2) caging chemistry for multi-wavelength visible light regulation of promoter function. Ultimately this technique will be used to elucidate the neuronal substrates of diverse mental states, such as fear, hunger, depression, anxiety, and addiction, thereby advancing the exploration of critical brain networks. This high-risk, high-reward project comprises multiple innovative features. Elements of the nascent reporter system described here, such as promoter strength, mechanism for regulating gene expression, choice of activating ligand, caging chemistry, and in vivo ligand and light delivery represent starting point that will benefit from extensive optimization. Once existing reporter components have been sufficiently refined, we envision replacing fluorescent reporters with recombinases, so that actuators can be expressed in identified cells for testing neuronal function. Other features, including the development of novel caged ligands, as well as additional methods for brain-wide activity reporting will also be addressed following achievement of our Aims. Despite the inherent risks, we are confident that our proposed system represents a fundamental and much-needed departure from existing techniques. We believe that our approach will evolve from its present status as a promising endeavor into a widely-used tool with the support of the BRAIN Initiative.